Robust Adaptive Fault-tolerant Asymptotic Tracking Control for Magnetic Levitation System Based on Nussbaum Gain and Neural Network

This paper studies a new fault-tolerant tracking control for magnetic levitation (MagLev) systems. Remarkably, the controlled system not merely admits unknown functions, but also allows unknown control directions. Under the backstepping framework, the neural network (NN) is constructively framed to estimate the unknown function in the MagLev system. Besides, to avoid high derivatives in the backstepping method, adaptive nonlinear filtering is applied to calculate the virtual control signal. Then, the Nussbaum gain technique is adopted to overcome the unknown control direction problem. Meanwhile, adaptive law in the proposed method is exploited to compensate for the influence of actuator failures. It turns out that the proposed adaptive fault-tolerant controller has the capability of guaranteeing the boundedness of all signals in the closed-loop system while steering the tracking error to converge to zero. Simulation analysis demonstrates the effectiveness of the proposed method.